Using R to Orchestrate APIs

Using R to Orchestrate APIs

A presentation for Research Data at the Edge, Day One of Duke Research Computing Symposium

Hosted by the Data & Visualization Services Department.

Presentation materials composed in Rmarkdown using Rstudio, stored in a Github Repository, Served via Github Pages.

• github Repo – https://github.com/libjohn/r-api-json
• Slides – https://libjohn.github.com/rcs2017/slides.html
• Notebook – http://libjohn.github.io/rcs2017/notebook.html

Outline

• API
• JSON
• R / RStudio

Why?

The Web has lots of stuff

• frontier beyond curated datasets
• stuff is wrapped in HTML
• HTML is transported over HTTP but composed for h2m consumption
• Intellectual Property rights bear serious consideration

API

Application Program Interface

• Built for machine-to-machine interactions
• Instructions for programs

---

Client / Server

• Make [R] interface with the web
• Same as h2m but now m2m

Human Simulation

A dramatization…

• Person uses Web Client

  • Person enters a URL
    

  • client & server negotiate
    
  • Information is sent back in wrapped HTML

  • Web Browser parses the HTML

m2m – development

dramatization: confused about the protocol

JSON

• Javascript Object Notation is a language-independent data format
• Currently the most common data data format for asynchronous client/server communication format
• Consists of key-value pairs

# from https://en.wikipedia.org/wiki/JSON
{
  "firstName": "John",
  "lastName": "Smith",
  "isAlive": true,
  "age": 25,
  "address": {
    "streetAddress": "21 2nd Street",
    "city": "New York",
    "state": "NY",
    "postalCode": "10021-3100"
  },
  "phoneNumbers": [
    {
      "type": "home",
      "number": "212 555-1234"
    },
    {
      "type": "office",
      "number": "646 555-4567"
    },
    {
      "type": "mobile",
      "number": "123 456-7890"
    }
  ],
  "children": [],
  "spouse": null
}

Example

To Follow Along

1. Open an RStudio Docker Container - https://vm-manage.oit.duke.edu/containers/rstudio
2. Project > New Project
3. Version Contrl > Git
4. Repository URL = https://github.com/libjohn/r-api-json.git > Create Project
5. GoTo Line 150-ish (“### Demonstration”) of the API-JSON-SYMPOSIUM.Rmd file

OMDB api

• http://www.omdb.org/
  • like http://imdb.com/
• no API keys requried
• http://www.omdbapi.com/

---

Demonstration

library(jsonlite)
# https://cran.r-project.org/web/packages/jsonlite/vignettes/json-aaquickstart.html
# for building tibbles
library(tidyverse)

Single JSON array

When the server response is a single JSON array, JSONlite makes viewing the data pretty simple.

oneJSONresult <- fromJSON("http://www.omdbapi.com/?t=rocky&y=&plot=full&r=json")

Let’s see the results in the next slide

---

oneJSONresult

$Title
[1] "Rocky"

$Year
[1] "1976"

$Rated
[1] "PG"

$Released
[1] "03 Dec 1976"

$Runtime
[1] "120 min"

$Genre
[1] "Drama, Sport"

$Director
[1] "John G. Avildsen"

$Writer
[1] "Sylvester Stallone"

$Actors
[1] "Sylvester Stallone, Talia Shire, Burt Young, Carl Weathers"

$Plot
[1] "Rocky Balboa is a struggling boxer trying to make the big time, working as a debt collector for a pittance. When heavyweight champion Apollo Creed visits Philadelphia, his managers want to set up an exhibition match between Creed and a struggling boxer, touting the fight as a chance for a \"nobody\" to become a \"somebody\". The match is supposed to be easily won by Creed, but someone forgot to tell Rocky, who sees this as his only shot at the big time."

$Language
[1] "English"

$Country
[1] "USA"

$Awards
[1] "Won 3 Oscars. Another 16 wins & 21 nominations."

$Poster
[1] "https://images-na.ssl-images-amazon.com/images/M/MV5BMTY5MDMzODUyOF5BMl5BanBnXkFtZTcwMTQ3NTMyNA@@._V1_SX300.jpg"

$Metascore
[1] "N/A"

$imdbRating
[1] "8.1"

$imdbVotes
[1] "387,927"

$imdbID
[1] "tt0075148"

$Type
[1] "movie"

$Response
[1] "True"

---

The vector object behaves as you would expect in R.

• You can list all the variable names.

names(oneJSONresult)

 [1] "Title"      "Year"       "Rated"      "Released"   "Runtime"    "Genre"      "Director"   "Writer"     "Actors"    
[10] "Plot"       "Language"   "Country"    "Awards"     "Poster"     "Metascore"  "imdbRating" "imdbVotes"  "imdbID"    
[19] "Type"       "Response"  

• List an individual element

oneJSONresult$Title

[1] "Rocky"

oneJSONresult$Awards

[1] "Won 3 Oscars. Another 16 wins & 21 nominations."

---

A JSON Matrix

The results of this code-snippet react differently between the console, the Notebook script (console), and the Notebook HTML output. In the Notebook script-output you can find the component name, in this case dollar-search: $Search. Or, you can use bracket notation: [[1]]. Once you identify the component name, it’s easier to identify the element names.

jsonSeriesResutlsMatrix <- fromJSON("http://www.omdbapi.com/?s=rocky&type=series&r=json&page=1")
jsonSeriesResutlsMatrix

$Search

$totalResults
[1] "20"

$Response
[1] "True"

---

Call the search results and coerce the JSON array into a data frame.

jsonSeriesResutlsMatrix$Search

---

jsonSeriesResutlsMatrix$Search$Title

 [1] "Rocky and His Friends"         "Dr. Jeff: Rocky Mountain Vet"  "Rocky Jones, Space Ranger"     "Rocky Mountain Law"           
 [5] "Rocky King, Detective"         "Rocky Road"                    "Rocky Mountain Bounty Hunters" "Rocky + Drago"                
 [9] "Rocky Point"                   "Rocky Star"                   

R Packages – Related

People who use JSONlite also use…

• httr – calls JSONlite in service to major goal of managing HTTP
• rvest – used for html parsing

Resources

• RStudio httR video
• JSONlite package
• listof images
• Movies of 1976
  • OMDB Top Movies
  • IMDB Most Popular